ARIMA model fitting and forecasting implementation

Question

I am new to time series modelling and I am trying to build a simple time series model using ARIMA methodology and forecast. I could write an R program to do the same, but I am more interested in writing my own arima(1,1,0) and predict() functions that R provides in Java / C++ / Python. While I can go about reading the theory behind ARIMA modeling and then implementing it, I still face a learning curve there, due to time constraints, I would like to take an approach where I can see a time series data sample, and a write up which walks me through the steps to compute arima() and predict() so that I can implement an algorithm. If someone can point me to a sourcecode / psedocode / step-by-step example via data, it would really be helpful.

Last time I went on the same cruise I found books on ARIMA rather than articles :) — Farhan Ahmed, May 02 '13 at 22:57
Well R is open source, so you already have some source code. — Rob Hyndman, May 03 '13 at 03:14

score 2 · Answer 1 · answered Nov 29 '13 at 10:29

If you're interested in fitting ARIMAs in Python then have a look at the statsmodels implementation. It already includes fitting, in-sample prediction, and forecasting methods. If you really want to write the function yourself then the source code might be useful, just as Farhan's C++ link could be for that language.

score 1 · Accepted Answer · edited Jan 21 '20 at 08:28

1

Here's an open source ARIMA implementation. It's under GNU license. Not sure if they would mind you learning from it!

It's an ARIMA implementation in C++, windows executable and source available. I think going through it you should be able to learn ARIMA the quick way.

edited Jan 21 '20 at 08:28

Dinei

103
4

answered May 02 '13 at 23:01

Farhan Ahmed

126

1

Welcome to CV. Would you mind giving a summary of what's there so readers can decide if it's what they're looking for, & / or in case of linkrot? – gung - Reinstate Monica May 03 '13 at 03:27
I hope that is sufficient :) – Farhan Ahmed May 03 '13 at 03:54
Thanks Rob, Farhan. If you have any suggestions on books for time series modeling for beginners, which also gives some background on basic concepts in statistics, that would be helpful. @gung - I'll need to cover quite some distance before I can give a decent summary of it. But I will try. – Mavendew May 12 '13 at 19:30
@Mavadu, my comment was to Farhan, not you. Note that he did provide a summary of what's at the link, as I had asked. If you want to know some books for studying time series, you could read this thread: books-for-self-studying-time-series-analysis (which was easily discoverable by searching the site--you might find more useful info by searching & reading). – gung - Reinstate Monica May 12 '13 at 20:21
@gung - Thanks I'll also make use of the link above! – Farhan Ahmed May 12 '13 at 22:31
@FarhanAhmed, the link was for Mavadu. – gung - Reinstate Monica May 12 '13 at 23:20
@Gung - does that mean it's illegal for me to click on it? :D – Farhan Ahmed May 13 '13 at 00:12
No, I hereby give you permission to click the link. – gung - Reinstate Monica May 13 '13 at 00:13

score 0 · Answer 3 · answered Jan 30 '22 at 14:40

There is imo no better material for studying ARIMA models than the textbook "Time series analysis: forecasting and control" by Box, Jenkins, Reinsel, Ljung. Box and Jenkins are major contributors to the development of the ARIMA. Chapters on forecasting and parameter estimation include pretty detailed explanations of how to implement algorithms.

ARIMA model fitting and forecasting implementation

3 Answers3